12466 research outputs found
Sort by
Design of Secure and Robust Cognitive System for Malware Detection
No full textThe computer systems for decades have been threatened by various types of hardware and software attacks of which Malware have been one of the pivotal issues. This malware has the ability to steal, destroy, contaminate, gain unintended access, or even disrupt the entire system. There have been techniques to detect malware by performing static and dynamic analysis of malware files, but, stealthy malware has circumvented the static analysis method and for dynamic analysis, there have been previous works that propose different methods to detect malware. However, these techniques do not perform well on stealthy malware. Moreover, the rising trend and advancements in machine learning has resulted into its numerous applications in the field of computer vision, pattern recognition to providing security to hardware devices. Machine learning based malware detection techniques rely on grayscale images of malware and tends to classify malware based on the distribution of textures in grayscale images. Albeit the advancement and promising results shown by machine learning techniques, attackers can exploit the vulnerabilities by generating adversarial samples. Adversarial samples are generated by intelligently crafting and adding perturbations to the input samples. There exists majority of the software based adversarial attacks and defenses. To defend against the adversaries, the existing malware detection based on machine learning and grayscale images needs a preprocessing for the adversarial data. This can cause an additional overhead and can prolong the real-time malware detection. So, as an alternative to this, we explore RRAM (Resistant Random Access Memory) based defense against adversaries. Therefore, the aim of this thesis is to address the above mentioned critical system security issues. The above mentioned challenges are addressed by demonstrating proposed techniques to design a secure and robust cognitive system. First, a novel technique to detect stealthy malware is proposed.The technique uses malware binary images and then extract different features from the same and then employ different ML-classifiers on the dataset thus obtained. Results demonstrate that this technique is successful in differentiating classes of malware based on the features extracted. Secondly, I demonstrate the effects of adversarial attacks on a reconfigurable RRAM-neuromorphic architecture with different learning algorithms and device characteristics. I also propose an integrated solution for mitigating the effects of the adversarial attack using the reconfigurable RRAM architecture
Teacher Self-Efficacy during COVID-19: A Qualitative Study on Experienced High School Teachers
Get PDFIn March of 2020, every teacher's career changed when the impact of COVID-19 prevented students from coming to school in-person. With the entire structure of education forced to change because of schools moving to online learning, Bandura's research on self-efficacy can provide insight to how individual teachers' self-efficacy beliefs adapted during COVID. This qualitative study focused on individual teacher's experiences involving the four sources that are known to impact self-efficacy, including mastery experiences, vicarious experiences, social persuasion, and physiological state. Given that research has found educators at the mid-point of their career are typically at their highest levels of teacher self-efficacy beliefs, seven teachers with at least ten or more years of experience in a public high school were interviewed for this research. This study observed that each teacher was impacted to varying degrees in each of the four sources of self-efficacy over the two years, between March 2020 and June 2022. The most notable self-efficacy impacts among all participants were with mastery experiences, social persuasion, and personal physiological state, as many teachers struggled with the loss of student interaction while teaching online. This study's findings suggest that the amount of support from schools and leadership played a significant role in teachers recovering their self-efficacy beliefs during COVID. In addition, results found that teachers who had an adaptable growth mindset toward teaching were more likely to easily rebound from self-efficacy loss. Based on these findings, school districts should evaluate what kinds of support they provide to teachers, both pre- and post-COVID. Focusing on what teachers need most, considerations should potentially address emotional adaptability and wellness opportunities to help teachers not only recover from the lingering effects of COVID, but also the everyday changes they face in the classroom
AI-Enhanced Software Vulnerability and Security Patch Analysis
Get PDFWith the increasing popularity of open-source software (OSS), their embedded vulnerabilities have been widely propagating to downstream software. Although timely applying security patches is the best practice to fix vulnerabilities, OSS users are hard to distinguish and prioritize security patches over tons of non-security patches (i.e., bug fixes, feature updates, etc.). Even worse, software vendors may silently release security patches without providing any explicit advisories. While users are unaware of security patches, attackers can still carefully inspect the patch code changes to exploit unpatched software. Therefore, automatically detecting security patches becomes imperative for software maintenance. In this dissertation, I describe my research efforts to address the above problems. First, I introduce an empirical study that reveals the insecure behavior of software vendors during maintenance and discloses the existence of silent security patches. Second, I present PatchDB, the first large-scale real-world patch dataset, that enables the training of data-hungry AI models for patch detection and facilitates future vulnerability/patch analysis research. An unsupervised method is developed to efficiently collect security patch samples from a huge number of unlabeled GitHub commits. Third, I present GraphSPD, a novel graph learning-based approach for automated security patch detection. By combining rich semantic properties of both pre-patch code and post-patch code in a joint graph structure and adopting a tailored multi-attributed graph convolution network to adapt diverse attributes in a patch graph representation, GraphSPD demonstrates state-of-the-art performance and detects 88 new silent security patches in popular GitHub projects
A Space of Our Own: An Intersectional Analysis of Black Queer Voices and Experiences
Get PDFThis auto-ethnographic study investigates the question of “Why do we gather?” where “we” refers to Black, queer, gender minority populations. Specifically, this research is focused on understanding why these individuals seek out similar others with whom to congregate and “hold space.” Twelve individuals, located in the DC, Maryland, and Virginia area, who attended one of three separate virtual meeting groups agreed to participate in one-on-one interviews and comprised the study population. Concepts of Blackness, queerness, and gender are specifically explored here as a means to better understand how they come together to inform the experiences of individuals whose lives are lived at the intersection of all three marginalized identities and who seek spaces in which those identities are shared. Using the lens of intersectionality and Foucault’s concept of heterotopia to analyze these interviews created a further focus on identity, gathering, and community. This work demonstrates how societal rejection can be exceedingly damaging to these individuals in ways that lead them to seek out others with similar experiences to create a space of safety
Quarter 1 2023
Get PDFThis issue of GEWEX Quarterly contains the articles on the following: a commentary reviewing the 30 Years of GEWEX BAMS paper; new GEWEX Panel and SSG members; YESS involvement in PAUWES, SRI Congress 2023, and the WCRP OSC; AGU H3S’s planning for new seminar series, 2023 AGU Fall Meeting events, and a fresh new look for its website; a new GEWEX initiative to improve groundwater modeling; LS4P examines the remote effect of Tibetan Plateau spring temperature and identifies high mountain land temperature as a possible first order source of S2S precipitation predictability (Phase I Highlights and Phase II Initiation); INARCH strives to increase insights into alpine cold regions hydrological processes; the Demistify project uses single-column model and large-eddy simulation intercomparisons to delve into radiation fog; PannEx holds 6th meeting to work towards a better understanding of Earth system components and their interactions in the Pannonian Basin; and a workshop on the role of early career researchers in contributing to Earth observations and geospatial science in Africa examines data access and availability, technology and technique handling, and collaboration
Project 1: Racial & Sex Disparities in Lupus & Transplant Outcomes
Get PDFBackground: Systemic lupus erythematosus (SLE) is an autoimmune disease with both high morbidity and mortality rates [1]. Clinical symptoms, laboratory findings, and optional biopsy results are the basis for early diagnosis of SLE. The disease is well-known for its butterfly rash that appears mainly on women at child bearing ages, and patients should be treated with antimalarials with hydroxychloroquine presenting the best results; it has a higher rate of remission, fewer relapses and reduced damage in the course of the disease [2]. Although there has been significant progress made in regards to SLE, the number of SLE patients who face End Stage Renal Disease (ESRD) has increased from 1.6 to 4.9 million from 1982 to 2004 [3]. Patients with ESRD were tested to predict a 3-year allograft survival after transplantation. Graft has an 84% chance of 3-year survival if the patient didn’t need dialysis during the first week compared to 58% if dialysis was needed during the first post transplantation week. For those who had needed dialysis for the first week, if their maintenance immunosuppressive regimen contained prednisone, took hemodialysis, and was older than 50 years, their graft had a lower probability of survival [1]. Adding on, much advancement and findings regarding SLE has led to the discovery of correlation between the disease and minority populations. The Centers for Disease Control and Prevention National Lupus Registries found that people of color (African Americans, APIs, and Hispanics) are more susceptible to SLE. Studies show that psychosocial stressors are potential exacerbators of SLE which can activate the inflammatory pathways. Often exposed to racial discrimination, people of color experience a large source of stress which can cause many health disparities, which in turn, can lead to the further worsening of SLE [4]. It has also been studied that after a kidney transplant, many patients have gotten rid of their Lupus while others, specifically African Americans (AA) have gotten Lupus back with other symptoms [3]. Adding on, five years after a kidney transplant for Lupus Nephritis (LN) graft survival for AF is at 63%, while for non-AA the graft survival rate is at 78.3%. Further, AA of lower median household income (MHI) are more likely to experience quicker graft loss compared to AA of higher MHI. For non-AA, there was no significant trend between MHI and graft loss [5]. While LN significantly increased the risk of graft failure, rejection remained the primary driver of graft failure incidence. These results emphasized that these findings should not discourage patients with Lupus from considering a kidney transplant [6].
Objective: The study investigated how the outcomes of kidney transplants differ by sex and race in the transplant recipients with SLE diagnosis
Dynamic Large-scale Screening Under Uncertainty
Get PDFIn this dissertation, we focus on large-scale infectious disease screening in dynamic and highly variable environments. Large-scale screening is crucial to any well-functioning healthcare system as it has a direct impact on human life as was demonstrated during the COVID-19 pandemic. Problems that arise in this context are often sequential as they involve taking screening actions, observing outcomes (which are often affected by uncertainty), and then adjusting decision accordingly. Given the nature of the problems involved, decisions that are more conservative (as opposed to ones that rely on average-case performance) are desirable. We focus on two aspects of the problem. The first deals with large-scale screening of established infections in which disease evolution can be predicted with a fairly good level of accuracy. In this setting, screening policies do not have a substantial impact on disease prevalence in the short-to-medium term and thus interactions between screening and populations characteristics can be relaxed. The second aspect focuses on large-scale screening during emerging pandemics. In this setting, modeling the interaction between screening decisions and disease dynamics on the short-term is crucial and thus must be explicitly incorporated in the decision-making framework. This unfortunately complicates the problem from a modeling perspective. For both problem aspects, we propose efficient algorithms and present case studies in the context of large-scale screening of donated blood in blood centers and of COVID-19 in the U.S
Perceived Agency Changes Trust in Robots
Get PDFThis thesis experimentally investigated the effect of people’s perception of a robot’s compliance (and resistance) to social norms on their evaluation of a robot’s perceived agency and trust. Participants reported higher perceived agency and trust to a norm-conforming robot compared to a norm-violating robot (experiment 1). Furthermore, results from experiment 1 found that perceived agency, regardless of how much a robot followed norms, was positively correlated with trust; therefore, suggesting that as people see a robot as having agency, they trust it more. While specifically examining the effect of negative attitudes on the relationship between social norms, perceived agency and trust, results from experiment 2 replicated experiment 1 and found that participants’ negative attitudes towards robots did not impact perceptions of trust in robots. These findings are useful in that they provide understanding for situations when people decide whether to trust a norm-conforming robot with perceived agency, regardless of if they feel negatively about robots. These results suggest that robot designers should pay careful attention on how to integrate robots with perceived agency cues. This is particularly important for safety critical robots in health care or the military because the more perceived agency a robot has, the more people seem to trust it (i.e., agency cues may not be representative of competency or ethics). Ideally, robot designers should understand the implications of integrating a robot high in perceived agency and low in capability, or conversely, low in perceived agency and high in capability. Future HRI research is necessary with specific emphasis on executable tools for measuring perceived agency for a diversified set of robots
Integrated Multi-stage Decision-Support for Enhanced Infrastructure Restoration under Uncertainty
Get PDFCritical infrastructure systems (e.g., water, transportation, and communication) provide fundamental services to communities. In recent decades, an increasing number of extreme events made infrastructure systems vulnerable and consequently brought severe impacts on community lifelines. To reduce such impacts, efficient infrastructure restoration is highly desired. The two determining phases of infrastructure restoration are the planning phase and the execution phase. In the planning phase, rapid infrastructure damage identification and restoration scheduling are critical to ensuring the efficient execution of restoration operations. In the execution phase, continuous restoration progress monitoring and control is needed for a timely identification of issues. Successful completion of these tasks involves emergency response agencies performing restoration operations under uncertain situations. However, during infrastructure restoration, coordination and communication among these agencies are challenging mainly because they have different roles, responsibilities, and jurisdictions within which they operate. In addition, unanticipated disruptions, changes in restoration tasks, and shifts in response demands further complicate the restoration process. With advanced data collection technologies, an increasing number of data sources (e.g., physical infrastructure information, socioeconomic information, and geographic information) have become available. Although being useful, transforming these data into useful decision support to facilitate rapid infrastructure restoration remains challenging. First of all, due to the unanticipated disruptions and the blocked road access, a valid and comprehensive damage inspection takes time to perform. Thus, only a limited amount of damage inspection data are available immediately following disruptions. Secondly, socially vulnerable communities are less prepared for disruptions. Therefore, it is important to restore the components that have greater impacts on socially vulnerable communities. Furthermore, to adapt to unanticipated disruptions that arise during the actual restoration, agencies need to update plans about task prioritization and resource allocation. As a result, the actual restoration progress deviates from the planned one. Last but not least, large-scale infrastructure damage typically spans across geographical and jurisdictional boundaries. Restoring these damaged components requires collaborative restoration efforts among various agencies, which takes significant communication and coordination efforts. To enable shared situational awareness and facilitate rapid infrastructure restoration, this research aims (1) to integrate geospatial correlation for addressing the data sparsity issue during infrastructure damage identification; (2) to prioritize restoration tasks while considering the socially vulnerable community demand; (3) to quantify the dynamic change of restoration progress during the restoration execution phase; and (4) to create a synchronized integration of various infrastructure restoration stages (damage identification, restoration scheduling, and progress monitoring). This research contributes to the domain of post-disaster infrastructure management by (1) proposing a systematic geospatial correlation-integrated approach for providing a quick spatial estimate of infrastructure damage status with incomplete information; (2) designing an equity-centered restoration scheduling approach that prioritizes restoration tasks while considering community social vulnerability; (3) performing real-time forecasting of infrastructure restoration progress and incorporate the associated uncertainties using Bayesian inference and earned schedule; and (4) establishing a framework that synchronizes various restoration stages. In practice, this research facilitates rapid infrastructure restoration by (1) providing a quick spatial estimate of infrastructure damage status, which greatly alleviates the effort and cost associated with field inspections; (2) generating up-to-date infrastructure restoration progress forecasting, which enables a timely observation of deviations between the actual and planned restoration progresses; (3) automatically recommending restoration task ranking while incorporating the socially vulnerable community demand, which could potentially alleviate the widening of the pre-existing socioeconomic disparities; and (4) promoting quick and shared restoration situational awareness among the involved emergency response agencies, which facilitates communication and coordination and helps overcome challenges resulting from fragmented restoration efforts
“ALWAYS WITH KINDNESS”: THE LIFE AND CAREER OF MUSIC EDUCATOR AND CONDUCTOR, ANTHONY MAIELLO
Get PDFAnthony Maiello is considered to be one of the most prolific conductors and music educators in America. He has done extensive work with bands and orchestras throughout his life and has impacted students and teachers worldwide as an in-demand clinician and guest conductor and an author of two prominent conducting textbooks. Maiello’s work has contributed to the standardization of preparedness on the podium through his adherence to regular, disciplined, and thorough score study, and he is considered by many a role model for music educators everywhere. This dissertation is the oral history of Maiello’s life and career. It includes stories, life lessons, and anecdotes gleaned from his 59 years of teaching, as well as his philosophies and acclaimed practices and processes as an instrumental music educator, conductor, and pedagogue. The purpose of this research was to create a principal source about Anthony Maiello, which will serve as a source of crucial information for music educators and conductors by providing insight for more mindful and impactful teaching, both now and in the future